Image sensors utilize a color filter array situated on a pixel array to filter the impinging light according to wavelength. Conventional color filters use a 2×2 pattern of red, green, and blue filters, where the green filters account for 50% of the total area, the red filters account for 25% of the total area, and the blue filters account for 25% of the total area. In order to obtain a full-color image, various demosaicing algorithms and color correction may be utilized to interpolate a set of color values for each pixel.
The spectral sensitivity of an image sensor is the relative efficiency of light detection as a function of wavelength. As pixel size decreases, however, crosstalk negatively influences the signal, as wavelengths intended for one pixel are captured in neighboring pixels and contributing to the wrong signal. Crosstalk diminishes the color signal of the affected color channels and increases the overlap in spectral responses of the different color channels.
To compensate for the increased overlap in spectral responses, intensive color correction can reproduce the diminished color gamut, but may also amplify noise. This leads to overall reduced signal-to-noise ratio (SNR) performance. The color correction matrices suited for image sensors with increased crosstalk sacrifice either color reproduction accuracy or SNR or both. As such, in low-light conditions, the pixel signals are already low and at risk of low SNR, so correcting for full color reproduction further reduces SNR. In bright-light conditions, the pixel signals have ample SNR and the increased noise from full color correction is less consequential.